CN107800644A - Dynamically configurable pipelined token bucket speed limiting method and device - Google Patents
Dynamically configurable pipelined token bucket speed limiting method and device Download PDFInfo
- Publication number
- CN107800644A CN107800644A CN201711089803.9A CN201711089803A CN107800644A CN 107800644 A CN107800644 A CN 107800644A CN 201711089803 A CN201711089803 A CN 201711089803A CN 107800644 A CN107800644 A CN 107800644A
- Authority
- CN
- China
- Prior art keywords
- token
- ram
- bucket
- message
- management module
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/215—Flow control; Congestion control using token-bucket
Abstract
The invention relates to a dynamically configurable pipelined token bucket speed limiting method and device. The invention can realize different speed limit requirements through dynamic configuration of parameters, simultaneously uses a dual-port RAM (random Access memory) to respectively store relevant information required by a token bucket algorithm, and forms streamlined reading and writing through dynamic configuration of time granularity and token increment granularity, thereby being suitable for different speed limit requirements of numerous streams; and the pipelined token management mode obviously improves the efficiency, thereby realizing the speed limit of large-scale data flow.
Description
Technical field
The present invention relates to network communication technology field, more particularly to it is a kind of can dynamic configuration streamlined limiting speed of token bucket side
Method and device.
Background technology
In available data network, message flow burst often causes network congestion, and the speed of message is entered in receiving terminal
Row limitation is a kind of method for solving its network congestion, and even the speed of message is less than defined speed and then normally received, otherwise
By packet loss or reduce its priority.
It is a kind of common method to message rate-limiting using token bucket algorithm.As shown in figure 1, system with defined speed to order
Token is added in board bucket, until token bucket completely then overflows；By the token ratio in message length and token bucket if having message arrival
Compared with allowing message if token abundance normal through while subtracting token corresponding to message length；Otherwise by packet loss or drop
Low priority is sent.
In above-mentioned token bucket algorithm, system adds a number of token into token bucket every certain time interval,
The token number often added by time interval or every time into token bucket takes fixed value.If time interval is taken into fixed value, its is big
The small frequency determined to token bucket addition token, smaller, system need to continually add token, and load is heavier, and therefore, it is difficult to reality
The limiting speed of token bucket of existing large-scale data stream；Larger, system needs the long period just to add a token, larger for speed
Message, after the token of token bucket is quickly run out of, until next time again add token before message can abandon or drop always
Level, the rate fluctuation in turn resulting in message are larger.It is smaller if the token number added every time to token bucket is taken into fixed value, it can prop up
The maximum speed limit held is smaller；The precision of larger then speed limit is relatively low, and precision can drastically reduce as the change of speed limit is big.
The method of another kind addition token is then with message to the trigger condition for being used as addition token.When a message arrives
Come, the time interval that system-computed current message arrives with a upper message, then by this time interval and token adding rate phase
Multiply the token number for now being needed to add to token bucket.Such a mode add token time interval depend entirely on message it
Between time interval, and many factors such as the length of the time interval between message and message, speed have it is related, if between message
Time interval it is long, beyond system timing scope, or token adding rate is changed in this time interval, then all can be right
The token number of addition causes larger error.
The content of the invention
To solve the above problems, the present invention provides one based on FPGA (Field Programmable Gate Array)
Kind can dynamic configuration streamlined limiting speed of token bucket method and apparatus, this method can pass through parameter for different speed limit requirements
Dynamic configuration is realized, while is stored respectively needed for token bucket algorithm using two-port RAM (Random Access Memory)
Relevant information, to form streamlined read-write, the efficiency of management tokens is improved, and then realize the speed limit of large-scale data stream.
To achieve the above object, implementation method of the present invention is as follows：
1) time granularity and token are increased into granularity as the parameter that can configure, according to requirement of each bar stream for speed limit by
Outside calculates laggard Mobile state configuration.Time granularity is the time interval for adding token；Token increase granularity is every time to token
The token quantity added in bucket；
2) token number in parameter, timer and the token bucket of above-mentioned dynamic configuration is stored respectively using two-port RAM, so as to
Independently it is written and read in each, avoids the access conflict of three kinds of information, and then realizes that full streamlined accesses；
3) relevant information of a plurality of stream is stored in a ram cell, then a plurality of stream can be accessed simultaneously by reading and writing a ram cell
Information, further improve token management efficiency.The ram cell represents memory space corresponding to an address ram；
4) timing is realized by counting of the counter to system clock cycle, respectively the parameter of streamlined cyclic access configuration
Compared with counter：When being more than or equal to its time granularity between the counter institute timing for having stream in the ram cell accessed
When, then quantity should be added into its token bucket increases the token of granularity for its token, then now sends signal notice renewal
Token number in token bucket, then restarts timing；Otherwise token need not be added, continues timing；
5) when a message arrives, whether permitted by token number decision-making message of the export control module in its corresponding token bucket
Perhaps normal forwarding；If token is inadequate, by packet loss or degradation；If token is sufficient, normally E-Packet and should be from order
The token number (message length) of consumption is subtracted in board bucket, then now also send the token number in signal notice renewal token bucket.
The export control module is the logic whether decision-making message normally forwards；
6) token number in token bucket is updated：According to above-mentioned addition token or the corresponding token bucket of signal-obtaining of consumption token
In token number, then the token number plus addition or re-write after subtracting the token number of consumption；
7) token management module is implemented separately with handling the export control module of message, in order to expand later.The token
Management module is above-mentioned generation token and the logic for updating token.
The present invention can be realized for different speed limit requirements by the dynamic configuration of parameter, while use two-port RAM (Random
Access Memory) relevant information needed for token bucket algorithm is stored respectively, by increasing granularity to time granularity and token
Dynamic configuration, to form streamlined read-write, the different speed limits for being applicable numerous streams require；And the token management mode of streamlined
Efficiency is significantly improved, and then realizes the speed limit of large-scale data stream.
Brief description of the drawings
Fig. 1 is the basic functional principle figure of token bucket algorithm；
Fig. 2 is the token management function structure chart of limiting speed of token bucket device；
Fig. 3 is the export control function structure chart of limiting speed of token bucket device.
Embodiment
Below in conjunction with the accompanying drawings, the specific implementation of the present invention is described in further detail.
It is the structure of an above-mentioned token management module shown in Fig. 2, the realization mechanism of token management module is described in detail：
1) a token management module includes tri- two-port RAMs of P_RAM, CNT_RAM and TB_RAM.P_RAM is storage configuration
Parameter P RAM, CNT_RAM be memory counter C RAM, TB_RAM be storage token bucket in token number T RAM.Institute
It is d to state three RAM depth, i.e., address ram arrives d-1 for 0, and each ram cell includes the relevant information of m bar streams, then an order
Board management module includes the relevant information of m*d bar streams altogether.Herein by the token number in the parameter P of configuration, counter C and token bucket
T is each stored using different two-port RAMs, in order to which three kinds of RAM independently carry out streamlined access, if there is n order
Board management module is worked simultaneously, then the information that can access n*m bar streams simultaneously is handled, and is remarkably improved token management efficiency,
And then realize the speed limit processing of large-scale data stream.
2) one group of ram port for taking P_RAM is used for the outside dynamic configuration for carrying out parameter：The parameter P of configuration includes time grain
Degree and token increase granularity, time granularity herein are used in what a time interval inside counting device for adding token should add up
Count value represents, in order to which compared with counter, token increase granularity is the token quantity added every time to token bucket.It is described
One group of ram port is one group of interface for being used for read-write operation in two-port RAM.
3) streamlined generation token.One group of ram port streamlined poll for taking CNT_RAM reads CNT_RAM, while uses P_
Also streamlined poll reads P_RAM for RAM another group of port, and the address ram that both read is identical.Assuming that the RAM of its reading
Address is x-1, then the parameter P for the configuration that the ram cell for reading P_RAM includes m bar streams：Px1...Pxm, CNT_RAM's
Ram cell includes the counter C of identical m bar streams：Cx1...Cxm；Respectively by the parameter P of configuration time granularity P [:] with
Counter C is accordingly contrasted one by one：If counter C be more than or equal to P [:], then it should add number to corresponding token bucket
Measuring increases the token of granularity for token, while by corresponding counter O reset；Otherwise token need not be added (should add in other words
Quantity is 0 token), while corresponding counter is added one.According to this principle, if addition order in need in m comparing result
Board, then send plus pulse is to update the token number T in its token bucket, without sending if it need not all add token.In addition
Take the counter C after CNT_RAM another group of ram port write-in renewal.It is mono- that described plus pulse should include currently processed RAM
Token number (the add_inc that first address (x) should each add with its m bars stream1..., add_incm).Whole processing procedure is carried out
Streamlined operates, and address ram carries out cyclic polling by 0 to d-1.
4) one group of ram port for taking TB_RAM is updated processing to the token number T in token bucket.Above-mentioned plus pulse and output
What control module was sent, which subtract pulse, is stored in FIFO (First In First Out).What the export control module was sent subtracts arteries and veins
Punching is that the stream ID comprising message and the pulse of the token number (RT) of consumption that send are believed after decision-making forwards a message consumption token
Number.The stream ID is every unique number flowed of mark.
5) above-mentioned renewal token is to handle the operation for adding pulse or subtracting pulse, the token number T of renewal TB_RAM storages.It is specific real
Existing process is as follows：Priority treatment subtracts pulse, and reading TB_RAM according to the stream ID subtracted in pulse corresponds to ram cell, it is assumed that reading
Address ram is y, then it includes the token number T of m bar streams：Ty1...Tym, then according to stream ID wherein it is determined that the order of renewal
Board number T, TB_RAM is re-write after being subtracted the consumption token number in pulse, the token number T of other m-1 streams is constant；If
Without pulse is subtracted, then start to process adds pulse, and reading TB_RAM according to the address ram added in pulse corresponds to ram cell, it is assumed that RAM
Address is x, then it includes the token number T of m bar streams：Tx1...Txm, and add the token that also needs comprising m bar streams add in pulse
Number：add_inc1...add_incm, TB_RAM is re-write after being corresponded addition.Processing procedure is with alternately read-write RAM
Mode realize streamlined handle.
6) such scheme is used, the relevant parameter of token bucket then increases granularity by bucket depth (Bucket_Size), token
(TokenAdd_GRN), token granularity (Token_GRN), time granularity (Time_GRN), system clock cycle (Tclk) and
RAM depth (d) forms.Bucket depth represents storable highest number of tokens in token bucket；Token increase granularity is represented every time to token
The token quantity of addition in bucket；Token granularity represents that a token allows the message byte number (being usually 1Byte) passed through；And this
The time granularity at place is not the time interval of real token addition, and it represents above-mentioned in the time interval of a token addition
The count value that counter should add up, such scheme are counted using streamlined cyclic polling RAM, then the addition of real token when
Between at intervals of Time_GRN*Tclk*d.Then the speed limit R of a certain bar stream calculation formula is as follows：
7) in above-mentioned implementation, Bucket_Size, Token_GRN, TclkBe fixed value with d, and TokenAdd_GRN with
Time_GRN be can dynamic configuration parameter, then R be represented by：R=C*TokenAdd_GRN/Time_GRN, C herein are
Bucket_Size、Token_GRN、TclkThe constant determined with d.So can be by changing TokenAdd_GRN's and Time_GRN
Ratio changes R.
8) for it is above-mentioned time interval or token granularity are taken into fixed value when the contradiction that occurs, this programme is to TokenAdd_
GRN and Time_GRN can dynamic configuration, and TokenAdd_GRN and Time_GRN occurrence can be according to each bar stream to R, essence
Degree is calculated with requirements such as acceptable bandwidth fluctuations and determined.Such as：Time_GRN is bigger, and precision is higher, but larger for speed
Stream, the time interval more conference for adding token causes its bandwidth fluctuation bigger；The bandwidth fluctuation is：It is larger for speed
Stream, after adding token, message can quickly run out of token, and message below can use without token, then can not normally E-Packet,
Add token again until next time, such bandwidth can be using the time interval that token adds as cyclic swing.
Fig. 3 show the export control function structure chart of limiting speed of token bucket device, describes n token management module while works
Situation.
1) each above-mentioned token management module includes a TB_RAM, and its TB_RAM one group of ram port has been used for renewal order
Board.And another group of ram port of the TB_RAM in n token management module is then all used for reading each bar to export control module
The token number T of stream.
2) caused token management module that subtracts pulse and may be sent in any one work after export control module transfer message,
So the interface for subtracting pulse is connected with n token management module respectively.
3) when export control module has message arrival, then transmitting message is started, while according to the stream ID of message it is determined that reading
TB_RAM, and read its corresponding ram cell.Based on above-mentioned token management module, it is assumed that the address ram of reading is z, then its
Corresponding ram cell contains the token number T of m bar streams：Tz1...Tzm, token number T grows with message according to corresponding to selecting stream ID
Spend PKT_LEN contrasts：If T>=PKT_LEN, then message normally forwards, while will include stream ID and the message length (order of consumption
Board number, RT) the pulse that subtracts issue corresponding token management module with for updating token；Otherwise it is that need to abandon or need by packet labeling
Degradation person's state, in order to carry out corresponding handle in message output logic below.
4) when the quantity increase for the data flow for needing speed limit, export control module only can be simply changed, and increase token management
The quantity of module, it is easy to extend.
Summary, it is proposed by the present invention can the streamlined limiting speed of token bucket method and apparatus of dynamic configuration pass through to time granularity
And the dynamic configuration of token increase granularity, the different speed limit requirements of applicable numerous streams；And the token management side of streamlined
Formula significantly improves efficiency, and then supports large-scale data stream speed limit.Obviously, those skilled in the art can be according to the present invention's
Technical concept carries out various changes and deformation to the present invention, and these modification and variation belong to the claims in the present invention and equivalent skill
Within the scope of art, then it should all belong to the protection domain of the claims in the present invention.
Claims (3)
1. it is a kind of can dynamic configuration streamlined limiting speed of token bucket method, matched somebody with somebody for different speed limit requirements by the dynamic of parameter
Realization is put, while stores the relevant information needed for token bucket algorithm respectively using two-port RAM, to form streamlined read-write, is carried
The efficiency of high management tokens, and then realize the speed limit of large-scale data stream, it is characterised in that implementation method is as follows：
1) time granularity and token are increased into granularity as the parameter that can configure, according to requirement of each bar stream for speed limit by
Outside calculates laggard Mobile state configuration, and time granularity is the time interval for adding token；Token increase granularity is every time to token
The token quantity added in bucket；
2) token number in parameter, timer and the token bucket of above-mentioned dynamic configuration is stored respectively using two-port RAM, so as to
Independently it is written and read in each, avoids the access conflict of three kinds of information, and then realizes that full streamlined accesses；
3) relevant information of a plurality of stream is stored in a ram cell, then a plurality of stream can be accessed simultaneously by reading and writing a ram cell
Information, further improves token management efficiency, and the ram cell represents memory space corresponding to an address ram；
4) timing is realized by counting of the counter to system clock cycle, respectively the parameter of streamlined cyclic access configuration
Compared with counter：When being more than or equal to its time granularity between the counter institute timing for having stream in the ram cell accessed
When, then quantity should be added into its token bucket increases the token of granularity for its token, then now sends signal notice renewal
Token number in token bucket, then restarts timing；Otherwise token need not be added, continues timing；
5) when a message arrives, whether permitted by token number decision-making message of the export control module in its corresponding token bucket
Perhaps normal forwarding；If token is inadequate, by packet loss or degradation；If token is sufficient, normally E-Packet and should be from order
The token number (message length) of consumption is subtracted in board bucket, then the token number in signal notice renewal token bucket is now also sent,
The export control module is the logic whether decision-making message normally forwards；
6) token number in token bucket is updated：According to above-mentioned addition token or the corresponding token bucket of signal-obtaining of consumption token
In token number, then the token number plus addition or re-write after subtracting the token number of consumption；
7) token management module is implemented separately with handling the export control module of message, in order to expand later, the token
Management module is above-mentioned generation token and the logic for updating token.
2. it is a kind of can dynamic configuration streamlined limiting speed of token bucket device, comprising token management module and export control module, its
It is characterised by, a token management module includes tri- two-port RAMs of P_RAM, CNT_RAM and TB_RAM,
The P_RAM is the parameter P of storage configuration RAM, and CNT_RAM is memory counter C RAM, and TB_RAM makes for storage
The RAM of token number T in board bucket；
Three RAM depth are d, i.e., address ram arrives d-1 for 0, and each ram cell includes the relevant information of m bar streams, then
One token management module includes the relevant information of m*d bar streams altogether, by the token in the parameter P of configuration, counter C and token bucket
Number T is each stored using different two-port RAMs, is easy to three kinds of RAM independently to carry out streamlined access, if there is n order
Board management module is worked simultaneously, then the information for accessing n*m bar streams simultaneously is handled, and improves token management efficiency, realizes big rule
The speed limit processing of mould data flow.
3. it is according to claim 2 it is a kind of can dynamic configuration streamlined limiting speed of token bucket device, it is characterised in that it is described
The specific implementation that n token management module works simultaneously is as follows：
1) each above-mentioned token management module includes a TB_RAM, and TB_RAM one group of ram port has been used for updating token,
And TB_RAM another group of ram port then all is used for reading the token number T of each bar stream to export control module；
2) caused token management module that subtracts pulse and can be sent in any one work, institute after export control module transfer message
It is connected respectively with n token management module with subtracting the interface of pulse；
3) when export control module has message arrival, then transmitting message is started, while according to the stream ID of message it is determined that reading
TB_RAM, and its corresponding ram cell is read, based on token management module, it is assumed that the address ram of reading is z, then it is corresponding
Ram cell contain the token number T of m bar streams：Tz1...Tzm, token number T and message length according to corresponding to selecting stream ID
PKT_LEN is contrasted：If T>=PKT_LEN, then message is normally forwarded, while the pulse that subtracts comprising stream ID and message length is issued
Corresponding token management module is with for updating token；Otherwise by packet labeling for need to abandon or need to the person's of degradation state, in order to
Message output logic below is handled corresponding to carrying out；
4) when the quantity increase for the data flow for needing speed limit, only simple modification export control module, and increase token management mould
The quantity of block, it is easy to extend.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711089803.9A CN107800644A (en) | 2017-11-08 | 2017-11-08 | Dynamically configurable pipelined token bucket speed limiting method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711089803.9A CN107800644A (en) | 2017-11-08 | 2017-11-08 | Dynamically configurable pipelined token bucket speed limiting method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107800644A true CN107800644A (en) | 2018-03-13 |
Family
ID=61549242
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711089803.9A Pending CN107800644A (en) | 2017-11-08 | 2017-11-08 | Dynamically configurable pipelined token bucket speed limiting method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107800644A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108399078A (en) * | 2018-03-19 | 2018-08-14 | 武汉斗鱼网络科技有限公司 | A kind of message frequency method for limiting, system, server and storage medium |
| CN108881055A (en) * | 2018-06-27 | 2018-11-23 | 深圳市风云实业有限公司 | token management method and device |
| CN110275670A (en) * | 2018-03-16 | 2019-09-24 | 华为技术有限公司 | Method, apparatus, storage equipment and the storage medium of data flow in control storage equipment |
| CN110569203A (en) * | 2019-08-09 | 2019-12-13 | 华为技术有限公司 | input control method and device and storage equipment |
| WO2020015356A1 (en) * | 2018-07-18 | 2020-01-23 | Huawei Technologies Co., Ltd. | Method and system for cluster rate limiting in a cloud computing system |
| US10771392B1 (en) | 2019-06-12 | 2020-09-08 | International Business Machines Corporation | Increasing efficiency to rate limiting |
| CN114567602A (en) * | 2022-03-08 | 2022-05-31 | 中国人民解放军国防科技大学 | Delay guarantee transmission system and method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080189522A1 (en) * | 2007-02-06 | 2008-08-07 | Ibm Corporation | Method and Apparatus for Enabling Resource Allocation Identification at the Instruction Level in a Processor System |
| CN102082693A (en) * | 2011-02-15 | 2011-06-01 | 中兴通讯股份有限公司 | Method and device for monitoring network traffic |
| CN102812670A (en) * | 2010-03-22 | 2012-12-05 | 飞思卡尔半导体公司 | Token bucket management apparatus and method of managing a token bucket |
| CN105024942A (en) * | 2015-05-29 | 2015-11-04 | 桂林电子科技大学 | Dynamic flow monitoring method |
-
2017
- 2017-11-08 CN CN201711089803.9A patent/CN107800644A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080189522A1 (en) * | 2007-02-06 | 2008-08-07 | Ibm Corporation | Method and Apparatus for Enabling Resource Allocation Identification at the Instruction Level in a Processor System |
| CN102812670A (en) * | 2010-03-22 | 2012-12-05 | 飞思卡尔半导体公司 | Token bucket management apparatus and method of managing a token bucket |
| CN102082693A (en) * | 2011-02-15 | 2011-06-01 | 中兴通讯股份有限公司 | Method and device for monitoring network traffic |
| CN105024942A (en) * | 2015-05-29 | 2015-11-04 | 桂林电子科技大学 | Dynamic flow monitoring method |
Non-Patent Citations (1)
| Title |
|---|
| 梅益波: "HINOC2.0_QOS优化设计及芯片级功能扩展", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110275670A (en) * | 2018-03-16 | 2019-09-24 | 华为技术有限公司 | Method, apparatus, storage equipment and the storage medium of data flow in control storage equipment |
| US11734183B2 (en) | 2018-03-16 | 2023-08-22 | Huawei Technologies Co., Ltd. | Method and apparatus for controlling data flow in storage device, storage device, and storage medium |
| CN108399078A (en) * | 2018-03-19 | 2018-08-14 | 武汉斗鱼网络科技有限公司 | A kind of message frequency method for limiting, system, server and storage medium |
| CN108399078B (en) * | 2018-03-19 | 2021-06-15 | 武汉斗鱼网络科技有限公司 | Message frequency limiting method, system, server and storage medium |
| CN108881055A (en) * | 2018-06-27 | 2018-11-23 | 深圳市风云实业有限公司 | token management method and device |
| WO2020015356A1 (en) * | 2018-07-18 | 2020-01-23 | Huawei Technologies Co., Ltd. | Method and system for cluster rate limiting in a cloud computing system |
| US10951532B2 (en) | 2018-07-18 | 2021-03-16 | Huawei Technologies Co., Ltd. | Method and system for cluster rate limiting in a cloud computing system |
| US10771392B1 (en) | 2019-06-12 | 2020-09-08 | International Business Machines Corporation | Increasing efficiency to rate limiting |
| CN110569203A (en) * | 2019-08-09 | 2019-12-13 | 华为技术有限公司 | input control method and device and storage equipment |
| CN110569203B (en) * | 2019-08-09 | 2021-11-30 | 华为技术有限公司 | Input control method and device and storage equipment |
| CN114567602A (en) * | 2022-03-08 | 2022-05-31 | 中国人民解放军国防科技大学 | Delay guarantee transmission system and method |
| CN114567602B (en) * | 2022-03-08 | 2023-09-05 | 中国人民解放军国防科技大学 | Delay guarantee transmission system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107800644A (en) | Dynamically configurable pipelined token bucket speed limiting method and device | |
| US7315550B2 (en) | Method and apparatus for shared buffer packet switching | |
| US10248350B2 (en) | Queue management method and apparatus | |
| CN103581055B (en) | The order-preserving method of message, flow scheduling chip and distributed memory system | |
| US7822915B2 (en) | Memory controller for packet applications | |
| CN101771598B (en) | Communication dispatching method of real-time Ethernet | |
| US10177997B1 (en) | Method and apparatus for packet and byte counting | |
| CN100546191C (en) | A kind of big capacity does not have the high-speed statistical counter that refreshes | |
| CN106789733B (en) | Device and method for improving large-scale network flow table searching efficiency | |
| CN106789697B (en) | method and device for improving aging efficiency of large-batch network flow tables | |
| CN105335323A (en) | Buffering device and method of data burst | |
| CN115695014A (en) | Access control list construction and data message processing method, device and system | |
| CN105681864B (en) | IP-based transport stream jitter removal method and device | |
| CN104407992B (en) | A kind of four port stores based on dual ported register array | |
| US7568074B1 (en) | Time based data storage for shared network memory switch | |
| CN110232029A (en) | The implementation method of DDR4 packet caching in a kind of FPGA based on index | |
| CN100499563C (en) | Increasing memory access efficiency for packet applications | |
| WO2023216777A1 (en) | Method and system for implementing asynchronous gating of large-scale time-sensitive networking | |
| Wang et al. | Block-based packet buffer with deterministic packet departures | |
| CN102025694A (en) | DSP (Digital Signal Processor) array based device and method for sending Ethernet data | |
| CN114153758B (en) | Cross-clock domain data processing method with frame counting function | |
| CN104683265A (en) | Accurate high-capacity packet counting method for 100G interfaces | |
| CN111126584B (en) | Data write-back system | |
| CN103384225A (en) | Flow control method based on input intersection cache fast packet switched network | |
| US6885591B2 (en) | Packet buffer circuit and method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180313 |